The ubiquitous isoprenoid biosynthetic pathway is responsible for the formation of the most chemically diverse family of metabolites found in nature (Hahn et al., J. Bacteriol. 178:619–624, 1996) including sterols (Popjak, Biochemical symposium no. 29 (T. W. Goodwin, ed.) Academic Press, New York, pp 17–37, 1970), carotenoids (Goodwin, Biochem. J. 123:293–329, 1971), dolichols (Matsuoka et al., J. Biol. Chem. 266:3464–3468, 1991), ubiquinones (Ashby and Edwards, J. Biol. Chem. 265:13157–13164, 1990), and prenylated proteins (Clarke, Annu. Rev. Biochem. 61:355–386, 1992). Biosynthesis of isopentenyl diphosphate (IPP), the essential 5-carbon isoprenoid precursor, occurs by two distinct compartmentalized routes in plants (Lange and Croteau, Proc. Natl. Acad. Sci. USA 96:13714–13719, 1999). In the plant cytoplasm, IPP is assembled from three molecules of acetyl coenzyme A by the well-characterized mevalonate pathway (Lange and Croteau, Proc. Natl. Acad. Sci. USA 96:13714–13719, 1999). However, a recently discovered mevalonate-independent pathway is responsible for the synthesis of IPP in plant chloroplasts (Lichtenthaler et al. FEBS Letters 400:271–274, 1997).
Following the synthesis of IPP via the mevalonate route, the carbon-carbon double bond must be isomerized to create the potent electrophile dimethylally diphosphate (DMAPP). This essential activation step, carried out by IPP isomerase, insures the existence of the two 5-carbon isomers, IPP and DMAPP, which must join together in the first of a series of head to tail condensation reactions to create the essential allylic diphosphates of the isoprenoid pathway (Hahn and Poulter, J. Biol. Chem. 270:11298–11303, 1995). Recently, it was reported that IPP isomerase activity was not essential in E. coli, one of many eubacteria containing only the non-mevalonate pathway for the synthesis of both 5-carbon isomers, suggesting the existence of two separate mevalonate-independent routes to IPP and DMAPP (Hahn et al., J. Bacteriol. 181:4499–4504, 1999). Thus, it is unclear whether an IPP isomerase is essential for the synthesis of isoprenoids in plant plastids as well. Regardless of whether IPP isomerase activity is present in plant plastids, the separation by compartmentalization of the two different biosynthetic routes, the mevalonate and deoxyxylulose phosphate pathways (or “non-mevalonate”), for IPP and DMAPP biosynthesis in plants is the fundamental tenet upon which the subject inventions are based.
The synthesis of IPP by the mevalonate pathway (Eisenreich et al., Chemistry and Biology 5:R221–R233, 1998) is cytoplasm based and occurs as follows: The condensation of two acetyl CoA molecules to yield acetoacetyl CoA is catalyzed by acetoacetyl CoA thiolase (EC 2.3.1.9). The addition of another molecule of acetyl CoA to acetoacetyl CoA is catalyzed by 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) synthase (EC 4.1.3.5) to yield HMG-CoA, which is reduced in the subsequent step to mevalonate by HMG-CoA reductase (EC 1.1.1.34). Mevalonate is phosphorylated by mevalonate kinase (EC 2.7.1.36) to yield phosphomevalonate, which is phosphorylated, by phosphomevalonate kinase (EC 2.7.4.2) to form mevalonate diphosphate. The conversion of mevalonate diphosphate to IPP with the concomitant release of CO2 is catalyzed by mevalonate diphosphate decarboxylase (EC 4.1.1.33).
In organisms utilizing the deoxyxylulose phosphate pathway (aka “non-mevalonate pathway”, “methylerythritol phosphate (MEP) pathway”, and “Rohmer pathway”), the five carbon atoms in the basic isoprenoid unit are derived from pyruvate and D-glyceraldehyde phosphate (GAP) (Eisenreich et al., 1998). Thus, synthesis of IPP and/or DMAPP by the non-mevalonate route, which occurs in plastids, is as follows: Pyruvate and GAP are condensed to give 1-deoxy-D-xylulose 5-phosphate (DXP) by DXP synthase (Sprenger et al., Proc. Natl. Acad. Sci. USA 94:12857–12862, 1997). The rearrangement and reduction of DXP to form 2-C-methylerythritol 4-phosphate (MEP), the first committed intermediate in the non-mevalonate pathway for biosynthesis of isoprenoids is catalyzed by DXP reductoisomerase (Kuzuyama et al., Tetrahedron Lett. 39:4509–4512, 1998). MEP is then appended to CTP to form 4-(cytidine 5′-diphospho)-2-C-methyl-D-erythritol (Rohdich et al., Proc. Natl. Acad. Sci. USA 96:11758–11763, 1999), followed by phosphorylation of the C2 hydroxyl group (Lüttgen et al., Proc. Natl. Acad. Sci. USA 97:1062–1067, 2000) and elimination of CMP, to form a 2,4-cyclic diphosphate (Herz et al., Proc. Natl. Acad. Sci. USA 97:2486–2490, 2000). Interestingly, Herz et al. reported the possible existence of bifunctional proteins with both YgbP and YgbB activities. Once the remaining steps to the fundamental five-carbon isoprenoid building blocks, IPP and DMAPP, in the non-mevalonate pathway are discovered, they will serve as additional targets for inhibitors with antiobiotic and herbicidal activity.
Since the non-mevalonate pathway is ultimately responsible for the biosynthesis of compounds critical for photosynthesis such as the prenyl side-chain of chlorophylls, which serve as lipophillic anchors for the photoreceptors and the photoprotective carotenoid pigments, any enzyme, gene, or regulatory sequence involved in the biosynthesis of IPP and/or DMAPP can be a potential target for herbicides. For example, the antibiotic fosmidomycin, a specific inhibitor of the enzyme DXP reductoisomerase (Kuzuyama et al., Tetrahedron Lett. 39:7913–7916, 1998) has been shown to have significant herbicidal activity, especially in combination with other herbicides (Kamuro et al. “Herbicide” U.S. Pat. No. 4,846,872; issued Jul. 11, 1989). The report of an Arabidopsis thaliana albino mutant being characterized as a disruption of the CLA1 gene, later revealed as encoding DXP synthase by Rohmer et al. (Lois et al., Proc. Natl. Acad. Sci. USA 95:2105–2110, 1998), also illustrates the potential of non-mevalonate pathway enzymes as targets for compounds with herbicidal activity. Accordingly, one of ordinary skill in the art can readily understand that as additional compounds are discovered exhibiting herbicidal activity based on their effects on the non-mevalonate pathway, those compounds could be used in accord with the teachings herein.
The synthesis of carotenoids from IPP and DMAPP takes place in plant plastids by a genetically- and enzymatically-defined pathway (Cunningham and Gantt, Ann. Rev. Plant Mol. Biol. 39:475–502, 1998). Enhanced production of carotenoids such as lycopene and β-carotene in plants is highly desirable due to the reported health benefits of their consumption (Kajiwara et al., Biochem. J. 324:421–426, 1997). Enhanced carotenoid production in plants can also have a dramatic effect on their coloration and be highly desirable to the growers of ornamentals, for example. The IPP isomerase reaction is considered to be a rate-limiting step for isoprenoid biosynthesis (Ramos-Valdivia et al, Nat. Prod. Rep. 6:591–603, 1997). Kajiwara et al. reported that the expression of heterologous IPP isomerase genes in a strain of E. coli specifically engineered to produce carotenoids resulted in over a 2-fold increase in β-carotene formation. Recently, it has been reported that expression of an additional gene for DXP synthase in an E. coli strain specifically engineered to produce carotenoids also increased the level of lycopene substantially (Harker and Bramley, FEBS Letters 448:115–119, 1999). Increased isoprenoid production also has been shown in bacteria by combining carotenogenic genes from bacteria with an orf encoding IPP isomerase; and was even further enhanced when additionally combined with the dxs gene from the MEP pathway to supply the precursors IPP and DMAPP (Albrecht et al. Nature Biotechnology 18: 843–846, 2000).
Accumulation of one specific isoprenoid, such as beta-carotene (yellow-orange) or astaxanthin (red-orange), can serve to enhance flower color or nutriceutical composition depending if the host is cultivated as an ornamental or as an output crop; and if the product accumulates in the tissue of interest (i.e. flower parts or harvestable tissue). In plants, tissue with intrinsic carotenoid enzymes can accumulate ketocarotenoids such as astaxanthin in chromoplasts of reproductive tissues of tobacco by addition of the biosynthetic enzyme beta-carotene ketolase (Mann et al., Nature Biotechnology 18: 888–892, 2000). Astaxanthin is the main carotenoid pigment found in aquatic animals; in microalgae it accumulates in the Chlorophyta such as in species of Haematococcus and Chlamydomonas. Thus, an increase in the essential 5-carbon precursors, IPP and DMAPP, by expression of orfs encoding IPP isomerase and orfs upstream thereof, can feed into the production output of such valuable isoprenoids in organisms other than bacteria.
As a further example of utility, Petunia flower color is usually due to the presence of modified cyanidin and delphinidin anthocyanin pigments to produce shades in red to blue groupings. Recently produced yellow seed-propagated multiflora and grandiflora petunias obtain their coloration from the presence of beta-carotene, lutein and zeaxanthin carotenoid pigments in combination with colorless flavonols (Nielsen and Bloor, Scienia Hort. 71: 257–266, 1997). Industry still lacks bright yellow and orange clonally propagated trailing petunias. Metabolic engineering of the carotenoid pathway is desired to introduce these colors in this popular potted and bedding plant.
Plant genetic engineering has evolved since the 1980s from arbitrarily located monocistronic insertions into a nuclear chromosome, often subject to multiple copies, rearrangements and methylation, to predetermined sites for defined multicistronic or multigenic operon insertions into a plastid chromosome (plastome), which thus far is thought impervious to typical nuclear gene inactivation. While breeding of crop plants by nuclear genome engineering is nevertheless a proven technology for major agronomic crops and for traits such as herbicide resistance, introgression of genes into the plastome is a highly promising breeding approach for several reasons as described by Bock and Hagemann (Bock and Hagemann, Prog. Bot. 61:76–90, 2000). Of note is the containment of transgenes in the transplastomic plant: Plastids are inherited through the maternal parent in most plant species and thus plastid-encoded transgenes are unable to spread in pollen to non-target species. Therefore plastid engineering can minimize negative impacts of genetically engineered plants. A report on potential transfer by pollen of herbicide resistance into weedy relatives of cultivated crops (Keeler et al., Herbicide Resistant Crops: Agricultural, Economic, Environmental, Regulatory and Technological Aspects, pp. 303–330, 1996) underscores the value of using plastid engineering rather than nuclear engineering for critical production traits such as herbicide resistance. Daniell et al. have recently demonstrated herbicide resistance through genetic engineering of the chloroplast genome (Daniell et al., Nat. Biotechnol., 16:345–348, 1998).
Moreover, plastids are the site of essential biosynthetic activity. Although most associate photosynthesis as the primary function of the chloroplast, studies document that the chloroplast is the center of activity for functions involving carbon metabolism, nitrogen metabolism, sulfur metabolism, biochemical regulation, and various essential biosynthetic pathways including amino acid, vitamin, and phytohormone biosynthesis. Crop traits of interest such as nutritional enhancement require genetic manipulations that impact plastid biosynthetic pathways such as carotenoid production. While nuclear-encoded gene products can be exported from the engineered nucleus into the plastid for such manipulations, the biosynthetic genes themselves can be inserted into the plastid for expression and activity. As we begin to pyramid multiple genes often required for pathway manipulations (such as the aforementioned carotenoid biosynthesis) the repeated use of selection markers is expected to lead to unstable crops through homology-dependent gene silencing (Meyer and Saedler, Ann. Rev. Plant. Physiol. Mol. Biol. 47:23–48, 1996). In addition, the requirement for higher expression levels of transgenes for effective phenotypes such as vitamin levels and herbicide and pest resistance levels often falls short in nuclear transformations. These deficiencies are overcome through plastid transformation or combining plastid with nuclear transformations: The plastid recognizes strings of genes linked together in multicistronic operons and, due to the high copy number of genes within a plastid and within plastids in a cell, can produce a hundred- to thousand-fold the amount of transgene product. Accordingly, there is a continuing need for improved methods of producing plants having transformed plastids (transplastomic plants).
Golden rice is one example for which plastid engineering can complement nuclear engineering of pathways that reside in the plastid, yet have met with limited success. The metabolic pathway for beta-carotene (pro-vitamin A) was assembled in rice plastids by introduction into the nuclear genome of four separate genes, three encoding plastid-targeted proteins using three distinct promoters, plus a fourth selectable marker gene using a repeated promoter (Ye et al. Science 287:303–305, 2000). The wild-type rice endosperm is free of carotenoids but it does produce geranylgeranyl diphosphate; combining phytoene synthase, phytoene desaturase, and lycopene-beta cyclase resulted in accumulation of beta-carotene to make “golden rice”. However, the quantity produced was lower than the minimum desired for addressing vitamin A deficiency. An increased supply of precursors for increasing intermediates, such as geranylgeranyl diphosphate, is predicted to significantly increase isoprenoid production. Insertion of an operon encoding the entire mevalonate pathway into the rice plastome of the “golden rice” genotype, using for example the methods as described in Khan and Maliga, Nature Biotechnology 17: 910–914, 1999, can provide a means for making improvements in metabolic engineering of this important monocot crop.
Proplastid and chloroplast genetic engineering have been shown to varying degrees of homoplasmy for several major agronomic crops including potato, rice, maize, soybean, grape, sweet potato, and tobacco including starting from non-green tissues. Non-lethal selection on antibiotics is used to proliferate cells containing plastids with antibiotic resistance genes. Plastid transformation methods use two plastid-DNA flanking sequences that recombine with plastid sequences to insert chimeric DNA into the spacer regions between functional genes of the plastome, as is established in the field (see Bock and Hagemann, Prog. Bot. 61:76–90, 2000, and Guda et al., Plant Cell Reports 19:257–262, 2000, and references therein).
Antibiotics such as spectinomycin, streptomycin, and kanamycin can shut down gene expression in chloroplasts by ribosome inactivation. These antibiotics bleach leaves and form white callus when tissue is put onto regeneration medium in their presence. The bacterial genes aadA and neo encode the enzymes aminoglycoside-3′-adenyltransferase and neomycin phosphotransferase, which inactivate these antibiotics, and can be used for positive selection of plastids engineered to express these genes. Polynucleotides of interest can be linked to the selectable genes and thus can be enriched by selection during the sorting out of engineered and non-engineered plastids. Consequently, cells with plastids engineered to contain genes for these enzymes (and linkages thereto) can overcome the effects of inhibitors in the plant cell culture medium and can proliferate, while cells lacking engineered plastids cannot proliferate. Similarly, plastids engineered with polynucleotides encoding enzymes from the mevalonate pathway to produce IPP from acetyl CoA in the presence of inhibitors of the non-mevalonate pathway can overcome otherwise inhibitory culture conditions. By utilizing the polynucleotides disclosed herein in accord with this invention, an inhibitor targeting the non-mevalonate pathway and its components can be used for selection purposes of transplastomic plants produced through currently available methods, or any future methods which become known for production of transplastomic plants, to contain and express said polynucleotides and any linked coding sequences of interest.
This selection process of the subject invention is unique in that it is the first selectable trait that acts by pathway complementation to overcome inhibitors. This is distinguished from the state of the art of selection by other antibiotics to which resistance is conferred by inactivation of the antibiotic itself, e.g. compound inactivation as for the aminoglyoside 3′-adenyltransferase gene or neo gene. This method avoids the occurrence of resistant escapes due to random insertion of the resistance gene into the nuclear genome or by spontaneous mutation of the ribosomal target of the antibiotic, as is known to occur in the state of the art. Moreover, this method requires the presence of an entire functioning mevalonate pathway in plastids. For example, if one of the enzyme activities of the mevalonate pathway is not present in the plastid, resistance will not be conferred.
There is strong evidence indicating that the origin of plastids within the cell occurred via endosymbiosis and that plastids are derived from cyanobacteria. As such, the genetic organization of the plastid is prokaryotic in nature (as opposed to the eukaryotic nuclear genome of the plant cell). The plastid chromosome ranges from roughly 110 to 150 Kb in size (196 for the green alga Chlamydomonas), much smaller than that of most cyanobacteria. However, many of the bacterium genes have either been lost because their function was no longer necessary for survival, or were transferred to the chromosomes of the nuclear genome. Most, but not all, of the genes remaining on the plastid chromosome function in either carbon metabolism or plastid genetics. However, many genes involved in these functions, as well as the many other functions and pathways intrinsic to plastid function, are also nuclear encoded, and the translated products are transported from the cytoplasm to the plastid. Studies have documented nuclear encoded genes with known activity in the plastid that are genetically more similar to homologous genes in bacteria rather than genes of the same organism with the same function but activity in the cytoplasm as reviewed for example in Martin et al. (1998) Nature 393:162–165 and references therein.
The process whereby genes are transported from the plastid to the nucleus has been addressed. Evidence indicates that copies of many plastid genes are found among nuclear chromosomes. For some of these, promoter regions and transit peptides (small stretches of DNA encoding peptides that direct polypeptides to the plastid) become associated with the gene that allows it to be transcribed, and the translated polypeptide relocated back into the plastid. Once this genetic apparatus has become established, the genes present in the plastid chromosome may begin to degrade until they are no longer functional, i.e., any such gene becomes a pseudogene.
As is common in prokaryotic systems, many genes that have a common function are organized into an operon. An operon is a cluster of contiguous genes transcribed from one promoter to give rise to a polycistron mRNA. Proteins from each gene in the polycistron are then translated. There are 18 operons in the plastid chromosome of tobacco (Nicotiana tabacum). Although many of these involve as few as two genes, some are large and include many genes. Evolutionary studies indicate that gene loss—as pseudogenes or completely missing sequences—occurs as individuals rather than as blocks of genes or transcriptional units. Thus other genes surrounding a pseudogene in a polycistronic operon remain functional.
The rpl23 operon consists of genes whose products are involved in protein translation. Most of these genes are ribosomal proteins functioning in either the large or small ribosomal subunit. One particular gene of note, infA, encodes an initiation factor protein that is important in initiating protein translation. Although this gene is functional in many plants, it is a pseudogene in tobacco and all other members of that family (Solanaceae), including the horticulturally valuable tomato, petunia, and potato crops. A recent survey of plant groups has indicated that there have been numerous loses of functionality of infA (Millen et al., Plant Cell 13: 645–658, 2001). This as well as other pseudogenes are identified in species whose chloroplast genomes have not yet been fully sequenced.
Pseudogenes such as infA become potential target sequences for insertion of intact orfs. Inserted orfs are controlled by regulatory upstream and downstream elements of the polycistron and are promoterless themselves. Pseudogenes are known for a multiplicity of crops and algae with chloroplast genomes that are already fully sequenced. Crops include grains such as rice and trees such as Pinus. Of note in the latter are the eleven ndh genes; all may serve as potential targets for transgene insertion.
Transplastomic solanaceous crops are highly desirable in order to eliminate the potential for gene transfer from engineered lines to wild species, as demonstrated in Lycopersicon (Dale, P. J. 1992. Spread of engineered genes to wild relatives. Plant Physiol. 100:13–15.). Amethod for plastid engineering that enables altered pigmentation, for improved nutrition in tomato or improved flower color in Petunia and ornamental tobacco as examples, is desirable for solanaceous crops. The infA gene is widely lost among rosids and some asterids; among the latter, infA is a pseudogene in all solanaceous species examined (representing 16 genera). The solanaceous infA DNA sequences show high similarity, with all nucleotide changes within infA being documented. Thus one set of flanking sequences of reasonable length as known in the art should serve for directed insertion of an individual or multiple orfs into the infA sites of the solanaceous species. It is documented in a solanaceous species that flanking sequences for genes to be inserted into the plastome are not required to be specific for the target species, as incompletely homologous plastid sequences are integrated at comparable frequencies (Kavanagh et al., Genetics 152:1111–1122, 1999).
The upstream 5′ region, often referred to as the 5′ UTR, is important on the expression level of a transcript as it is translated. Knowing the translation products of surrounding genes in a polycistron allows one to select a pseudogene site that is affiliated with a strong 5′ UTR for optimizing plastid expression in a particular tissue. The plastid genome in many plant species can have multiple pseudogenes that are located in different polycistronic sites. So, if one has a choice, one can select a site based on whether it is actively transcribed in green vs non-green plastid; and then if the polycistron has high or low relative expression in that plastid type. Moreover, monocistronic mRNA of ndhD was detected in developed leaves but not in greening or expanding leaves of barley (Hordeum vulgare), despite this gene being part of a polycistronic unit as reported by del Campo et al. (1997) Plant Physiol 114:748. Thus, one can time transgene product production by treating an inactive gene, based on developmental expression, as a pseudogene for targetting and integration purposes using the invention disclosed herein.
Algal species are becoming increasingly exploited as sources of nutraceuticals, pharmaceuticals, and lend themselves to aquaculture. Mass production of the isoprenoid compound astaxanthin produced by the green microalga Haemotcoccus is one successful example of the above. Metabolic engineering that would increase product yields and composition in microalgae would significantly benefit the industry. The development of organellar transformation for the unicellular green alga Chlamydomonas reinhardtii, with its single large chloroplast, opens the door for conducting studies on genetic manipulation of the isoprenoid pathway. Filamentous or multicellular algae are also of interest as untapped biofactories, as are other nongreen algae whose pathways for producing unique fatty acids, amino acids, and pigments can be ameliorated for commercial benefit.
The biolistic DNA delivery method is a general means with which to transform the chloroplast of algae (Boynton and Gillham, Methods Enzymol. 217:510–536, 1993). Sequencing of at least six plastomes from algae should facilitate transformation systems by confirming insertion sites, including pseudogene sites, and the regulatory elements directing heterologous gene expression. What is required is a dominant marker for selection of stable transformants to which natural resistance is absent (Stevens and Purton, J. Phycol 33: 713–722, 1997). For Chlamydomonas, chloroplasts can be engineered using markers that confer spectinomycin resistance following their integration into the plastome via homologous recombination. By utilizing the polynucleotides disclosed herein in accord with this invention, an inhibitor targeting the non-mevalonate pathway and its components can be used for selection purposes of transplastomic algae produced through currently available methods, or any future methods which become known for production of transplastomic algae, to contain and express said polynucleotides and any linked coding sequences of interest. This is a novel selection vehicle for transplastomic algae. Moreover, elevating the supply of essential precursors for isoprenoid production in algae as described above is enabled by this invention.